Task 3, Segmenting the original image

You can segment the original image using either the FANTASM plug-in program or the Fuzzy C-Means algorithm. The FANTASM plug-in program, a different version of the Fuzzy C-Means algorithm, provides segmentation that is less sensitive to noise. This section provides directions for both:
Using FANTASM to segment the image
Using the Fuzzy C-Means algorithm to segment the image

To use the FANTASM plug-in program

1 Select the original image.
2 Select Plug-In > Algorithms > FANTASM. The FANTASM dialog box opens.
Figure 1. FANTASM dialog box

3 Complete the dialog box:
In Number of desired classes, select either 2 classes if the CPF was stripped off the image previously or 3 classes if the CPF is still in the image.
In Segmentation, select Hard only, which is the only segmentation needed for measurements.
4 Click OK. The Initial Centroids dialog box opens.
Figure 2. Initial Centroids dialog box


5 Click OK.
A series of progress messages appear, and data appears on the Data page in the Output window (Figure 3).
Figure 46. Data on the Data page in the Output window


In a few moments, processing completes, and an image appears (Figure 4A).
Figure 4. FANTASM image


To use the Fuzzy C-Means algorithm
1 Select the original image.
2 Select Algorithms > Segmentation> Fuzzy C-Means > Single Channel in the MIPAV window.



The Fuzzy C-Means dialog box (Figure 5) opens.
Figure 5. Fuzzy C-Means dialog box


3 Complete the dialog box:
In Number of desired classes, select either 2 classes if the CPF was stripped off the image previously or 3 classes if the CPF is still in the image.
In Segmentation, select Hard only, which is the only segmentation needed for measurements.
4 Click OK. The Initial Centroids dialog box (Figure 6) opens.
Figure 6. Initial Centroids dialog box


5 Click OK. A series of progress messages appear. In a few moments when processing completes:
Data appears on the Data page in the Output window (Figure 7)
Figure 7. Data on the Data page in the Output window


An image appears (Figure 8).

Figure 8. Image produced by using Fuzzy C-Means algorithm


Task 4, Transforming Talairach image and Talairach VOIs-1 to the original image